Clinicians depend on electronic health records (EHRs) and other high-tech tools to give them the best data at the point of care, when they’re interfacing directly with patients and have to make potentially lifesaving decisions. However, as studies and tales of the last five years of computerized healthcare have revealed, there’s also such a thing as too much data.

Clinical decision support (CDS) tools have emerged as a solution to this issue by transforming the most critical patient data into actionable alerts. But while the idea of using CDS in tandem with EHRs piqued the interest of healthcare leaders, clinicians ended up flooded with far more information than they could possibly need, including details that have no real bearing on decision-making.

In order to truly leverage the benefits of EHRs, healthcare leaders need to adopt smarter health IT systems, especially smarter CDS tools. Perhaps the easiest way to understand the current need for smarter CDS tools is by examining how an environment of too much data has led to multiple issues — starting with alert fatigue.

When EHRs gained widespread adoption in the first few years of the Centers for Medicare and Medicaid (CMS) EHR Incentive Program, which rolled out in 2010, the term “alert fatigue” became a catchphrase. Though EHRs aren’t the only source of alert fatigue; it’s also fueled by multiple “smart” devices with alerting capabilities that began populating care settings around the same time.

In the seven years since “meaningful use” entered the healthcare lexicon, a growing number of studies have demonstrated the ways clinicians are inundated with a high number of alerts, many of which are more disruptive than helpful to workflow. One recent 31-day study^[1] conducted in an academic hospital found that 66 adult ICU beds generated more than 2 million alerts, or 187 warnings per patient per day. Many of these warnings were clinically inconsequential: For example, 88.8 percent of the 12,671 annotated arrhythmia alarms issued during the study period were false positives.

With such a large percentage of false alerts, it’s little wonder that clinicians become accustomed to ignoring them altogether – effectively negating any potential gains from these technologies. A 2013 report published in the Journal of the American Medical Informatics Association^[1] noted that “several studies cite very high override rates ranging between 49 percent and 96 percent.”

CDS systems joined the ranks of solutions designed to improve care in the post-meaningful use era, in this case issuing guidance alerts, based on analysis of patient data, to clinicians through EHR screens. Yet until recently, some CDS tools inadvertently contributed to alert fatigue. One 2012 study^[1], for example, noted a direct correlation between the response rates to the alerts and alert exposure.

After investing millions of dollars in their EHR, practice management and CDS systems, healthcare leaders are often befuddled on how to leverage all the data these technologies generate in the most impactful way.

Fortunately, technology has grown smarter over the last five years, and that’s good news for healthcare leaders who are feeling anxious over the domino effect of alert fatigue.

By utilizing smarter CDS tools and other technologies that rely on sophisticated “behind the scenes” algorithms, healthcare systems will reap the true benefits of technology, receiving critical medical information in real time, instead of a deluge of data noise.

The algorithms that bolster smarter CDS tools work by reading data in real time, as it is entered into an EHR or other system, searching for keywords, terms or other metrics that stand out among patient data. Once the CDS combs through the real-time data and extracts the most important information, it filters this information through existing, pre-set modules (e.g., for conditions such as heart disease or diabetes) and compares patient data to baseline measures. After all this work, a smart CDS generates and prioritizes alerts, which are calls to action.

So if, for example, a patient is admitted to an acute-care facility after suffering from chest pains, a smart CDS can read real-time biometric data (such as body temperature) and combine it with multiple ancillary sources of demographic data (such as age, medication allergies, etc.), and produce an appropriate alert, such as a warning a clinician not to administer a common medication that would cause anaphylactic shock.

A smart CDS is one that adheres to the “Five Rights” CDS principle^[1]: delivering the right information to the right people, through the right channels, in the right intervention formats, at the right points in workflow for optimized decision making. In addition, a smart CDS can operate in tandem with various EHRs, and coexist alongside most other health IT specifications, which will become more important as initiatives such as the Medicare Access and CHIP Reauthorization Act (MACRA) take hold. (As a side note, healthcare providers who choose the MIPS track under MACRA will need to use an EHR to adhere to guidelines in the “Advancing Care Information” category, which emphasizes information exchange and interoperability). Truly smart CDS tools are capable of mining data from multiple EHRs — bypassing HL7 connectivity concerns.

Ultimately, a smarter CDS will enable clinicians to make important choices at the point of care, such as whether to prescribe a certain medication or proceed with an alternate course of care if a patient isn’t responding to a particular protocol. Without smart technology, a CDS can’t support an organization properly by maximizing an EHR’s potential.

A smart CDS system is one that turns information delivery into an art form, by intuitively knowing what information clinical teams need, and when they need it – so clinical teams can do their best work and patients can make the biggest improvements.

^[1] https://psnet.ahrq.gov/resources/resource/28461

^[2] https://www.researchgate.net/profile/Jonathan_Teich/publication/231210831_Drug-drug_interactions_that_should_be_non-interruptive_in_order_to_reduce_alert_fatigue_in_electronic_health_records/links/53f392110cf2155be35098aa.pdf

^[3] https://psnet.ahrq.gov/resources/resource/24474

^[4] https://www.cms.gov/eHealth/downloads/ClinicalDecisionSupport_AlertsTipsheet.pdf